Professional Documents
Culture Documents
Supercritical fluid extraction of organic compounds from solids and aqueous solutions
selected pollutants such as polycyclic aromatic hydrocarbons and
chlorophenols from soil samples. A novel mechanically agitated
countercurrent continuous extraction column was introduced by extracting
ethanol and 1-butanol from aqueous solutions in the rotating disk column
and in the Oldshue-Rushton column.
Antero Laitinen
VTT Chemical Technology
Laboratory of Chemical Engineering and Plant Design
Department of Chemical Technology
Helsinki University of Technology
Abstract
In this thesis, the extraction of organic compounds from solid and liquid
matrices using supercritical carbon dioxide as a solvent was studied.
Dense carbon dioxide is a very good solvent for the cyclic amide, 6-caprolactam.
6-caprolactam exhibits relatively high solubilities, i.e. up to around 17 wt% at
ambient temperatures, and at pressures below 220 bar. Carbon dioxide can be
used to dissolve and extract 6-caprolactam from solid and liquid matrices.
3
high solvent-to-feed ratios. Column capacities for both types of column were
measured, and the total throughput calculated as the feed plus solvent flow rate
divided by the smallest cross-sectional area was 5080 m3h-1m-2. A mechanically
agitated column successfully combines the high capacity of supercritical spray
columns and the high efficiency of supercritical packed columns.
4
Preface
The research described in this thesis was carried out at the Chemical laboratory
of the Technical Research Centre of Finland (VTT) during the years 19931998.
I would like to thank VTT Chemical Technology for the excellent working
facilities and for providing me with the opportunity to test my ideas. Financial
assistance from Tekes is gratefully acknowledged.
My most sincere thanks are due to my parents, my sister and my brother, and
my friends for their unselfish support during this work.
Antero Laitinen
5
List of publications
The thesis is based on the following publications, which are referred to by
Roman numerals:
6
The author' s contribution
The author has had an active role in all stages of the work reported in this
thesis, including the design and construction of the experimental
equipment, experimental design, the development of the working
procedures, the carrying out of the experiments reported in publications,
the interpretation the results, and the writing of the papers.
Matti Jntti did the modeling work in Paper . Olli Aaltonen contributed
significantly to the writing of Paper . The work of Martti Alkio in
designing and constructing the apparatus used in the soil extraction
experiments is acknowledged. A lot of the practical work in Papers V, VI
and VII was carried out by Juha Kaunisto.
7
Contents
Abstract .................................................................................................................3
Preface...................................................................................................................5
List of publications................................................................................................6
Abbreviations ......................................................................................................10
1. Introduction...................................................................................................11
8
6.1.2 Column extraction procedure..........................................................38
6.1.3 Calculation of column efficiency....................................................39
6.2 Column extraction results ........................................................................41
6.2.1 Column efficiency...........................................................................41
6.2.2 Column capacity .............................................................................46
References ...........................................................................................................51
Papers IVII
9
Abbreviations
a interfacial area, m2/m3
C pollutant concentration in the soil, mg/kg
D column diameter, m
E extraction coefficient = mS/F
F flow rate of feed, kg/h
h holdup
HETS height equivalent to a theoretical stage, m
HTUod height of a transfer unit based on the dispersed phase, m
Kod overall mass transfer coefficient based on the dispersed phase, 1/s
m distribution coefficient (concentration in solvent phase/concentration
in feed phase)
NTUod number of transfer units based on the dispersed phase
p pressure, bar
S flow rate of the solvent, kg/h
t extraction time, min
T temperature, C
y weight fraction of solute in the dispersed phase, mole fraction in section 4
V superficial velocity, m/s
w water content in the soil, wt%
Vs slip velocity, m/s
Z total agitated height of the column, m
density, g/ml
Subscripts
Superscripts
* equilibrium value
through smallest cross-sectional area
10
1. Introduction
In recent years, the use of supercritical fluid extraction for the removal of
organic compounds from different liquid and solid matrices has attracted much
attention. This technique has some advantages over more conventional
separation techniques, largely due to the unique physical properties of
supercritical fluids. Supercritical fluids exhibit a liquid-like density, while their
viscosity and diffusivity remain between gas-like and liquid-like values. Also,
the recovery of a supercritical solvent after extraction can be carried out
relatively simply by reducing the pressure and evaporating the solvent, because a
supercritical solvent is usually a gas at NTP. There are several applications in
the food processing, pharmaceutical and chemical industries, notably in the
decaffeination of coffee beans and the production of polyethylene [1].
11
distillation; the solvent is then recycled. Extraction processes are applied in the
petroleum industry to separate close-boiling hydrocarbon mixtures (e.g. the BTX
-process), in the pharmaceutical industry to separate heat-sensitive antibiotics,
and in the chemical industry to separate, for example, caprolactam from aqueous
feed. Supercritical fluids have been tested as solvents in several extraction
applications, such as the extraction of ethanol or acetic acid from aqueous
solutions [1].
12
2. Supercritical fluids
A substance is at the supercritical state, i.e.a supercritical fluid, when its
temperature and pressure are simultaneously higher than the critical values. A
supercritical fluid exists in a single fluid phase possessing characteristics
between those of gases and liquids. Figure 1 illustrates a pressure-temperature
diagram in the critical region of pure carbon dioxide.
Above the critical temperature a liquid phase will not appear regardless of how
much the pressure is increased. The critical pressure is the pressure, which
causes the gas to become a liquid at the critical temperature. The density of a
compound at the critical point is called the critical density. The critical
temperatures, pressures and densities of some common solvents are listed in
Table 1.
13
Table 1. Critical properties of selected solvents [1, 3].
Tc pc c
SOLVENT (C) (bar) (g /ml)
Ammonia 132.5 113.5 0.24
Benzene 289.0 48.9 0.30
n-Butane 152.0 38.0 0.23
Carbon dioxide 31.1 73.8 0.45
Chlorotrifluoromethane 28.8 39.5 0.58
Dichlorodifluoromethane 111.7 39.9 0.56
Ethane 32.2 48.9 0.20
Ethanol 243.4 63.8 0.28
Ethylene 9.3 50.4 0.22
Isopropanol 235.3 47.6 0.27
Methanol 240.5 79.9 0.27
Nitrous oxide 36.5 72.3 0.46
n-Propane 96.8 42.6 0.22
Propylene 91.9 46.2 0.23
Toluene 318.6 41.1 0.29
Water 374.2 221.2 0.34
14
should remember, that to recycle the supercritical solvent, it must be compressed
again. This can be a significant cost factor, if the difference between the pressure
in the extraction vessel and the pressure in the separator is relatively large.
Even though the density of a supercritical fluid increases with pressure and
becomes liquid-like, the viscosity and diffusivity remain between liquid-like and
gas-like values [4]. Additionally, supercritical fluids exhibit almost zero surface
tension, which allows facile penetration into microporous materials. As a result
of the advantageous combination of physicochemical properties, the extraction
process can often be carried out more efficiently with a supercritical solvent than
it can with an organic liquid one. A comparison of selected physicochemical
properties of liquids, gases and supercritical fluids is presented in Table 2.
The critical temperature (Tc) of the compound depends on the polarity of the
compound. As seen in Table 1, the critical temperatures of non-polar gases, such
as carbon dioxide or ethane, are below 50 C, whereas for polar compounds,
such as methanol or water, the critical temperature is well above 200 C. In
practice, especially in food-related industries, it is usually desirable that the
critical temperature of the solvent is below 100 C. Therefore, the solvents
commonly used for supercritical operations are low molecular weight gases,
such as carbon dioxide, ethane and propane.
Carbon dioxide is the most commonly used solvent in industrial practice for
several reasons. Carbon dioxide has a technically convenient critical pressure
and temperature of 73.8 bar and 31.1 C, respectively. It is non-toxic, non-
15
flammable, non-reactive, non-corrosive, and abundant. Furthermore, it is the
second least expensive solvent after water [5], and it does not leave any solvent
residue after extraction.
16
3. Applications of supercritical fluid
extraction
3.1 Historical background
In 1879 Hannay and Hogard [7] reported that changes in pressure caused
inorganic salts like cobalt chloride to dissolve in or precipitate from ethanol at a
temperature above the critical temperature of ethanol. Later, Villard [8]
demonstrated that gases such as methane, ethylene, carbon dioxide and nitrous
oxide dissolve a number of liquid or solid compounds, such as camphor, stearic
acid or paraffin wax. Since then, several authors have described the ability of a
supercritical fluid to dissolve solid materials. The solubility of napthalene in
supercritical ethylene was studied by Diepen and Scheffer in 1948 [9], followed
by several other studies [1015] in which the solubility of napthalene in a variety
of supercritical solvents was examined. In 1954 Francis [16] presented a study,
where he reported phase diagrams for 464 ternary systems containing liquid
carbon dioxide, and measured the solubility of 261 organic compounds in near
critical liquid carbon dioxide. Even though supercritical carbon dioxide was not
applied in the study, the data can be used to assess the solvent power of
supercritical carbon dioxide, because a compound which is soluble in liquid
carbon dioxide is usually also soluble in supercritical carbon dioxide.
17
3.2 Food processing applications
Liquid carbon dioxide from the container is pressurized by the pump and heated
to the extraction temperature by the heat exchanger. Pressurized carbon dioxide
flows through the extraction vessel and dissolves the soluble extract. Carbon
dioxide loaded with the extract is then flashed in the pressure release valve and
carbon dioxide, now in the vapor and liquid phase, flows into the separator,
where the liquid fraction is vaporized by heating. The extract is not soluble in
18
carbon dioxide vapor and can be removed from the bottom of the separator.
Carbon dioxide vapor is condensed and returned to the container. A heat pump
system can be installed to transfer heat from the condenser to the separator.
Supercritical fluids serve as a reaction medium for the reactants, catalysts and
products, or they can be a solvent and a reactant at the same time, as in
polyethylene-free radical polymerization. Additionally, the properties of
supercritical fluids may be utilized to separate the product from the reactants or
catalyst by reducing the pressure of the reaction mixture. In the mid 1980s it was
19
discovered that enzymes are active in supercritical carbon dioxide [21, 22].
Since then, several supercritical fluid research groups have investigated
enzymatic reactions, and more than 200 papers have been published. The basic
motivation is to replace the hazardous organic solvents with environmentally
acceptable carbon dioxide. Also, enzymes are more stable in the non-aqueous
supercritical reaction environment than in the conventional aqueous
environment. Review articles concerning enzymatic catalysis in supercritical
fluids are available [2325].
Supercritical fluids can be used as particle formation media [32, 33]. Two
techniques have been used to make particles from supercritical fluids. Rapid
expansion from supercritical solutions (RESS) is used to form fine particles of
substances which are soluble in a supercritical solvent. In the RESS technique
the substance is first dissolved in a supercritical solvent, after which the pressure
20
is rapidly decreased in a specifically designed nozzle, resulting in
supersaturation of the substance in the supercritical solvent and leading to the
formation of small particles with a narrow particle size distribution. Gas anti-
solvent recrystallization (GAS) is used for substances, which are not soluble in a
supercritical solvent. The substance is dissolved in a liquid solvent such as
methanol, and a supercritical solvent acts as an anti-solvent to rapidly precipitate
the substance in a nozzle.
21
3.6 Separations in continuously operated columns
22
glyserides, hydrocarbons, citrus oils, organic compounds from aqueous extracts,
and others.
Tiegs et al. [51] have used countercurrent a packed extraction column equipped
with reflux to extract fatty acid ethyl esters from fish oil. Simes et al. [52]
reported mass transfer results for the separation of a mixture of two terpenes of
similar volatilities. Carmelo et al. [53] have designed a column for the extraction
of free fatty acids (FFA) from olive oil.
23
4. Solubility measurements
In the first paper of this thesis [I], the solubility of 6-caprolactam in supercritical
carbon dioxide was reported. The ability of dense carbon dioxide to dissolve 6-
caprolactam could possibly be utilized in the industry related to nylon
production. Carbon dioxide could be used as a solvent to separate 6-caprolactam
from various solid or liquid feeds in applications related to production, the
recycling of used nylon or the purification of waste streams containing 6-
caprolactam.
4.1 Experimental
24
N
TI PI
F E A
C
B
D
J
I
K L
G
H M
25
0.08
0.07
0.06
0.05
y
0.04
0.03
0.02
0.01
80 100 120 140 160 180 200 220 240
p, bar
26
Table 3. Mole fraction solubility of 6-caprolactam in supercritical carbon
dioxide [I, Table 2].
27
5. Soil extraction measurements
Prior to the experimental work, a literature review concerning the removal of
organic pollutants from soil was written [II]. The primary purpose of the soil
extraction work was to study the extraction of selected pollutants from soil-water
slurries. The basic idea was to avoid the technically cumbersome semi-batch
extraction process by first slurrying the soil with water, and then continuously
pumping the soil-water slurry to the extractor, where the pollutants were
removed from the soil.
5.1 Experimental
The characterization and preparation of the soil samples as well as the extraction
apparatus and working procedures are briefly described in this section.
EURO-soils 5 and 6 were chosen as standardized soils. These soils had been
dried in an air-conditioned laboratory for 3 to 4 months, sieved, homogenized,
and sterilized by radiation. They were received from the EEC Joint Research
Center (Ispra, Italy). Table 4 shows the main pedological parameters of the soils.
28
The EURO-soil samples were spiked using the following procedure.
Approximately 300 g of EURO-soil was weighed and put into a 1000 ml flask.
Because the soil samples were completely dry, 810 wt-% water was first added
to the soils. The flask was then rotated in a rotavapor (Heidolph VV 2000) for at
least 12 hours, in order to homogenize the soil moisture. The soil was then
spiked with trichlorophenol, 2,3,4,6-tetrachlorophenol and pentachlorophenol
(Fluga AG, tech.) and phenanthrene (Fluka AG, purum). The desired amount of
these compounds were weighed and dissolved into 200 ml dichloromethane
(Fluga AG, purum). The mixture was decanted into the flask containing the soil.
The soil slurry was then rotated for at least 12 hours, after which
dichloromethane was evaporated slowly to dryness at room temperature. The
spiked soil was then kept in the fume hood for 24 hours to assure that there was
no dichromethane residue in the soil. Soil samples taken from contaminated soil
areas were not characterized.
29
SPB-35 (Supelco). More details concerning the analytic procedure are given in
Paper [IV].
CO2
CONTAINER
In Papers III and IV the extraction studies were carried out using either
standardized soil, which was spiked with known amounts of pollutants before
extraction, or by using soil samples taken from contaminated soil areas
originally containing pollutants.
30
5.2.1 Extraction from spiked soil
The effect of soil moisture content on the residual pollutant concentration in the
soil after extraction was the most important factor to be studied [IV]. The effect
of extraction pressure, temperature, pH, carbon dioxide mass-flow rate, and
extraction time on the residual pollutant concentration in the soil after extraction
was also studied systematically [III, IV]. The experimental parameters and
results are presented in Paper IV (Tables 3 and 4). Table 5 illustrates the best
results and the set of extraction parameters with which these results were
obtained.
Soil Pollutant T p w pH S t Cf Ci
5
1 64 295 7.2 4.5 1.4 20 81 1 760
2 64 295 7.2 4.5 1.4 20 101 1 870
3 64 295 7.2 4.5 1.4 20 43 550
6
1 50 200 7.9 5.9 3 45 35 2 490
2 64 295 14.0 6.7 1.5 20 8.5 940
3 64 295 14.0 6.7 1.5 20 6.8 510
31
phenol is regarded as contaminated soil. In most extractions the final pollutant
concentrations in the soil after extraction were not below the recommended
values.
The initial water content in the soil tested in the experiments was 740 wt%. The
experiments were carried out at relatively low temperatures (2565 C), and the
extraction pressure range was 100300 bar. The soil pH was changed by adding
a weak aqueous solution of either H2SO4 or K2CO3 to the soil. By using a
relatively high solvent flow rate (0.63 kg/h), it was possible to keep the
extraction time relatively short (2045 min).
The effect of soil moisture content on the pollutant concentration in the soil after
extraction was dominant in all of the soil extraction experiments. Generally, the
extraction of pollutants from soil-water slurries was not successful. Yeo and
Akgerman [63] have affirmed this conclusion by extracting phenol and
naphthalene from soil-water slurries at 47 C and 175 bar. They were able to
extract only 15 % of the initial amount of naphthalene and 60% of the initial
amount of phenol from the soil.
It was found, that in order to reach the level of acceptable residual pollutant
concentrations in the soil, it was necessary to keep the soil moisture content
below approximately 10 wt%. It is possible that slurrying the soil with water
reinforces the adsorption between the soil and the hydrophobic pollutant, and the
extraction result weakens. According to Andrews [64], the addition of small
amounts of polar modifiers like water or methanol dramatically reduces PAH
adsorption, indicating competition for polar adsorption sites on the soil. The
highly polar aqueous environment deactivates the polar adsorption sites, leaving
weak inductive, dispersive, and hydrophobic interactions as the only mechanism
for sorption. However, non-ionic and nonpolar PAHs interact primarily with
organic soil matter. Due to the negligible solubility of PAHs in water, adsorption
onto the soil is highly favored in soil-water slurries [64]. It also seems
reasonable to assume that water forms an extra mass transfer barrier, filling the
soil pores, and carbon dioxide is not able to contact and dissolve the pollutant. A
third possibility is that carbon dioxide saturated with water is no longer a good
solvent for non-polar pollutants.
32
The effect of other parameters on the residual pollutant concentration in the soil
was moderate compared with the effect of soil moisture content, even though it
was noticed that the highest extraction pressure and temperature were favorable.
This result is obvious, because a higher extraction temperature typically
increases the pollutant desorption from the soil. Also, increasing the extraction
pressure raises the solvent density and therefore increases the solubility of the
pollutant in the supercritical carbon dioxide solvent.
Polycyclic aromatic hydrocarbons (PAHs) were extracted from railway yard soil
[IV]. The soil initially contained 50 mg/kg of PAHs. The soil was extracted with
supercritical carbon dioxide at 200 bar and 40 C. In the first experiment, the
soil moisture content was 10 wt%. In this case the residual concentration of
PAHs in the soil after extraction was 7 mg/kg. In the second experiment, the soil
was slurried in water, and the soil moisture content was 50 wt%. This
experiment was clearly worse. The residual concentration of PAHs in the soil
after extraction was 30 mg/kg. The concentration of PAHs in the soil as a
function of time is presented in Figure 6. Obviously, the presence of a relatively
large amount of water (50 wt%) did make the efficient extraction of PAHs from
the soil more difficult.
33
concentration in soil, mg/kg
100 moisture
50 wt%
20
10
5
0 20 40 60 80 100 120 140
time, min
Figure 6. Extraction of PAHs from railway yard soil with supercritical carbon
dioxide at 200 bar and 40 C.
Chlorophenols were extracted from sawmill soil [IV]. The soil contained
initially 62 mg/kg of 2,4,6-trichlorophenol, 3440 mg/kg of 2,3,4,6-
tetrachlorophenol, and 800 mg/kg of pentachlorophenol. In the first experiment,
the soil moisture content was 10 wt%. In this case the residual concentrations of
the three chlorophenols in the soil after extraction were 0.5 mg/kg for 2,4,6-
trichlorophenol, 23 mg/kg for 2,3,4,6-tetrachlorophenol, and 24 mg/kg for
pentachlorophenol. The concentration of 2,4,6-trichlorophenol in the soil as a
function of extraction time is presented in Figure 7.
34
Concentration in soil, mg/kg recovery-%
100 99,5 recovery
50
99 concentration
20
10 98,5
5
2 98
1
97,5
0,5
0,2 97
0 20 40 60 80 100 120 140
time, min
In the second experiment, the soil was slurried in water, and the soil moisture
content was 50 wt%. As in the case of the extraction of PAHs from the railway
yard soil, the second experiment was worse. The residual concentrations of the
three chlorophenols in the soil after extraction were 3.1 mg/kg, 320 mg/kg, and
230 mg/kg, respectively.
35
6. Continuous column extractions
In Papers V, VI and VII a mechanically agitated high-pressure column was
designed and tested by extracting organic compounds from aqueous solutions
using dense carbon dioxide as a solvent. In Paper V a mechanically agitated
continuous countercurrent rotating disk extraction column (RDC) was
introduced by extracting ethanol from an aqueous solution. In Paper VI an
Oldshue-Rushton type column was tested in the same application. In Paper VII
the extraction of 1-butanol from an aqueous solution was studied using
supercritical carbon dioxide solvent in an Oldshue-Rushton column.
6.1 Experimental
The equipment, working procedures and calculation methods that were used in
Papers VVII are presented in this section.
6.1.1. Equipment
36
Figure 8. Column extraction installation.
The column's internal design consists of removable parts that allow the use of
different agitator geometries or even different types of columns. The
characteristics of the RDC- and Oldshue-Rushton columns which were used in
the experiments [V, VI, VII] are summarized in Table 6.
Table 6. Rotating disk and Oldshue-Rushton column characteristics.
Character RDC Oldshue-Rushton
Height of the column, mm 2 000 2 000
Total agitated height of the column, mm 1 400 1 400
Column diameter, mm 35 35
Compartment height, mm 20 20
Stator plate opening diameter, mm 24 24
Smallest cross-sectional area, mm2 452 452
Number of compartments 99 99
Impeller or disk diameter, mm 20 20
Impeller or disk height, mm 4 1
Number of flat blades 4 -
37
A commercial high-pressure extraction unit (Nova Werke AG) was used to feed
carbon dioxide to the column. Carbon dioxide flow-rates were measured by two
high-pressure turbine flow meters (EG&G Flow Technology) in the solvent and
extract streams. No dispersers or other similar devices were used to create the
drops. The liquid feed stream was pumped to the top of the column by a high-
pressure piston pump (Lewa AG). The liquid feed and carbon dioxide solvent
streams were preheated to the extraction temperature by means of constant
temperature water baths.
An aqueous solution containing ethanol [V, VI] or 1-butanol [VII] was used as a
continuous phase and carbon dioxide as a dispersed phase in all of the
experiments. Prior to each run, the column was heated to the operating
temperature. The system was then pressurized to approximately 50 bars with
carbon dioxide, after which the aqueous solution was pumped to the top of the
column until the phase boundary was seen in the window at the top of the
column. The carbon dioxide compressor (Nova AG) was then turned on, the
system was pressurized to the extraction pressure, and the extract stream valve
was opened to let carbon dioxide flow through the continuous liquid phase. The
feed pump was then turned on again and the optical sensor controlling the liquid
level was activated. A tachometer was used to measure the speed of the rotor.
All of the experiments were carried out at 100 bar and 40 C. Each extraction
experiment lasted approximately 4 hours. Fresh carbon dioxide from the
container was used in all of the experiments to avoid using a preloaded solvent.
During the run the feed and raffinate tanks were weighed, and the carbon
38
dioxide flow rates were measured every 20 minutes. After running for 2 hours,
the samples were taken from the feed, raffinate and extract streams once every
20 minutes. The extract from the separator was continuously collected. The
ethanol concentration in all of the samples was analyzed by GC. More details
concerning the sampling and material balance deviation are presented in the
Papers V, VI and VII.
The method described by Rathkamp et al. [59] was used to measure the
dispersed phase holdup. The data was obtained by shutting the feed pump and
carbon dioxide compressor off at steady state conditions, and measuring the
time required for the carbon dioxide drops that entered the bottom of the column
to rise to the top of the column. This information combined with the flow rate of
the dispersed phase was used to calculate the volume of the dispersed phase in
the column, and the holdup for the run was calculated. The raising drops were
recorded at various flow rates and rotor speeds by a video-camera looking
through the window in the middle of the column. The average diameter of the
drop was calculated for each experiment from a video freeze-frame.
It was not possible to measure flooding points with the optical sensor
construction, because the optical sensor opened the raffinate stream valve at the
bottom of the column when the liquid level started to rise. However, when the
solvent and feed flow rates reached certain values, the raffinate stream valve was
opento such an extent that most of the solvent flow came out of the column with
the raffinate stream. At this point the column was no longer operable. This point
was used to describe the capacity of the column.
39
y1
dy
NTU od = (1)
y2
y y *
Z
HTU od = (2)
NTU od
Vd
Kod a = (3)
HTU od
ln E
HETS = HTU od (4)
E 1
where E is the extraction coefficient._ The slip velocity through the smallest
cross-sectional area of the column Vs describes the relative velocity of the
phases, and is defined as
_ _
_ V V
Vs = d + c (5)
h 1 h
40
_ _
where Vd and Vc are the relative velocities of the dispersed and continuous
phases through the smallest cross-sectional area, and h is the dispersed phase
holdup.
In Table 7 the values of the overall mass transfer coefficient Koda and the height
equivalent to a theoretical stage HETS are compiled [V, VI, VII]. The extraction
efficiencies in RDC- and Oldshue-Rushton columns are relatively similar to
each other. The measured Koda values for the extraction of ethanol from water in
the RDC-column range from 0.006 s-1 to 0.015 s-1, and in the Oldshue-Rushton
column the values range from 0.009 s-1 to 0.012 s-1. The measured HETS values
are between 0.8 m and 0.37 m. Slightly lower Koda values were measured for 1-
butanol extraction than in ethanol extraction. In the case of 1-butanol extraction,
the measured Koda values are 0.0020.003 s-1, and the HETS values are 0.40.5
m in the Oldshue-Rushton column.
In the literature the values of the overall mass transfer coefficient Koda measured
for spray and packed continuous countercurrent supercritical fluid extraction
columns generally range from 0.006 s-1 to 0.04 s-1, depending on the type of
column, packing and solvent-to-feed ratio [57, 59]. Values of the height
equivalent to a theoretical stage HETS for small-scale spray, packed and sieve
tray supercritical fluid extraction columns have been measured by other authors
[54, 56, 57, 59]. The reported HETS values for the supercritical spray, packed
and sieve tray columns are typically in the range of 0.20.5 m.
It can be concluded that the Koda and HETS values measured in Papers V,VI and
VII for the RDC- and Oldshue-Rushton columns are of the same order of
magnitude as those presented for small-scale supercritical spray and packed
columns in the literature.
41
Table 7. Mass transfer efficiencies in mechanically agitated columns [V, VI, VII].
42
Extraction of 1-butanol in the Oldshue-Rushton column
43
The effect of agitation on the values of the overall mass transfer coefficient Koda
and the height equivalent to a theoretical stage HETS at the highest experimental
solvent-to-feed ratios are presented in Figures 9 and 10.
0.02
0.015
Koda (1/s)
0.01
0.005
0
0 100 200 300 400 500
rotor speed (rpm )
44
1
0.9
0.8
HETS (m)
0.7
0.6
0.5
0.4
0.3
0 100 200 300 400 500
rotor speed (rpm )
45
offset by increased axial mixing, resulting in an optimum rotor speed for each
operating condition.
Mechanical agitation seems to have an effect on column efficiency only at
relatively high solvent-to-feed ratios, e.g. S/F = 13 or S/F = 14. One possible
explanation is that at relatively low solvent-to-feed ratios, equilibrium is
approached in the column and mechanical agitation can no longer enhance the
extraction.
6.2.2 Column capacity
The total throughput of the RDC- and Oldshue-Rushton columns calculated as
the volumetric feed plus solvent flow rate divided by the smallest cross-sectional
area of the column is presented in Figure 11. The total throughput of the RDC-
column was approximately 80 m3 h-1 m-2 without agitation. Agitation slightly
lowered the value to 74 m3 h-1 m-2 when the rotor speed was 300 rpm. Without
agitation the total throughput of the Oldshue-Rushton column was
approximately 70 m3 h-1 m-2, whereas at 300 rpm the total throughput was
approximately 50 m3 h-1 m-2.
100
90
total throughput (m/h m)
80
70
60
50
40
30
0 50 100 150 200 250 300 350
rotor speed (rpm )
Figure 11. Effect of agitation on column capacity at 40 C and 100 bar. ()
Oldshue-Rushton column; () RDC -column.
46
Without agitation the capacities of the RDC- and Oldshue-Rushton columns are
relatively similar. The relative deviation of the values is approximately 12 %. At
300 rpm the capacity of the RDC -column is almost the same as without
agitation, whereas the capacity of the Oldshue-Rushton column is reduced
significantly at 300 rpm. One possible explanation is that the Oldshue-Rushton
type impeller provided more intense agitation than the relatively small-diameter
RDC -disk. When the agitation power input is increased, the drop diameter
decreases, and consequently the slip velocity and the column capacity decrease
as well. At relatively low power inputs this effect is rather small, but at higher
power inputs the drop size reduction reaches the point at which it becomes so
large that the slip velocity and the capacity are substantially reduced.
The dispersed phase holdup and slip velocities were also measured in the RDC-
and Oldshue-Rushton columns. The results are presented in Table 8.
47
Table 8. Dispersed phase holdup and slip velocities in the RDC- and Oldshue-
Rushton columns.
The measured dispersed phase holdup values are in the 0.030.20 range. As
expected, all of the holdup values increase with increasing rotor speed. All of the
slip velocity values decrease with increasing rotor speed because of increasing
holdup.
48
7. Concluding remarks
Considering that there is a common need among chemical engineers to find
environmentally benign solvents, the use of relatively inexpensive carbon
dioxide solvent should increase in the future, provided that carbon dioxide
proves to be a technically feasible and economically competitive solvent in
selected industrial applications. In this thesis, the extraction of organic
compounds from solid and liquid matrices using supercritical carbon dioxide as
a solvent was studied.
Dense carbon dioxide is a very good solvent for the cyclic amide, 6-caprolactam,
and can be used to dissolve and extract it from solid and liquid matrices. The
ability of dense carbon dioxide to dissolve 6-caprolactam could possibly be
utilized in the industry producing nylon.
Certain criteria must be fulfilled before solvent extraction can be considered for
soil remediation [69]. The first criterion is that the process must be capable of
yielding uncontaminated residual material. The results of this thesis indicate that
at least 80-90 % of the initial amount of pollutant can be removed at moderate
temperatures even in the most difficult cases, and in many cases over 99 % of
the initial amount of pollutant can be removed from soil. Supercritical extraction
is best suited to sand or silt type soils, which have a low adsorption capacity.
The extraction results are found to be case-dependent and soil remediation at
moderate temperatures is observed to be difficult if the pollutant binds strongly
to the soil or if the acceptable level of residual concentration is very low. To
verify technical feasibility, a commercial-scale extraction system will have to be
successfully operated over an extended period. Only one such system is reported
to have been installed in Port Arthur, Texas [69]. Generally, there is plenty of
evidence in industry, e.g. the extraction of caffeine from coffee beans, that high-
pressure extraction is a viable technology. The solids handling in high-capacity,
high-pressure batch processes is a notable disadvantage, and an efficient method
of moving the soil into the high-pressure extractor would be a significant
advancement.
49
column. A special pressure-compensating sealing system was used to lead the
electrically driven agitator shaft through the column head. The capacity of a
mechanically agitated supercritical extraction column is typically higher than
that of a conventional liquid-liquid extraction column due to the fact that the
density difference of the phases in supercritical extraction columns is larger than
in conventional extraction processes. A mechanically agitated column
successfully combines the high capacity of supercritical spray extraction
columns and the high efficiency of packed supercritical extraction columns.
50
References
1. McHugh, M., Krukonis, V. Supercritical Fluid Extraction, 2nd ed.,
Butterworth-Heinemann, Boston, 1994.
51
11. Tsekhanskaya, Y. V., Iomtev, M. B., Mushkina, E. V. Solubility of
diphenylamine and naphthalene in carbon dioxide under pressure. Russ. J.
of Phys. Chem., 36(1962), pp. 1177.
52
21. Hammond, D. A., Karel, M., Klibanov, A. M., Krukonis, V. J. Enzymatic
reactions in supercritical gases. Appl. Biochem. Biotechnol., 11(1985),
pp. 393.
53
31. Hitzler, M. G., Poliakoff, M. Continuous hydrogenation of organic
compounds in supercritical fluids. Chem. Commun., (1997), pp. 1667.
38. Low, G., Duffy, G. J., Sharma, S. D., Chensee, M. D., Weir, S. W.,
Tibbett, A. R. Transportable supercritical fluid extractor unit for treating
of contaminated soils. In: Proceedings of the 3rd International Symposium
on Supercritical Fluids, 1719 October 1994, Strasbourg, France. Pp.
275280.
54
40. Ehrlich, R., Huang, C.-P. Remediation of soil contaminated by 2-
chlorophenol and 2,4,6-trichlorophenol using supercritical fluid
extraction. Hazard. Ind. Waste, 26(1994), pp. 472.
41. Aaltonen, O., Alkio, M., Lundell, J., Ruohonen, S., Parvinen, L.,
Suoninen, V. Polypeptide purification with industrial-scale supercritical
fluid chromatography. Pharmaceutical Tech. Europe, 18(1998), pp. 112.
43. Knez, Z., Posel, F., Sencar, P. Treatment of pesticide polluted water with
supercritical CO2. In: The 2nd International Symposium on Supercritical
Fluids, 2022 May 1991, Boston, USA. Pp. 777782.
45. Furuta, S., Ikawa, N., Fukuzato, R., Imanishi, N. Extraction of ethanol
from aqueous solutions using supercritical carbon dioxide. Kagaku
Kogaku Ronbunshu, 15(1989), pp. 519.
47. Randall, J. M., Schulz, W. G., Morgan, A. I. Extraction of fruit juices and
concentrated essences with liquid carbon dioxide. Confructa, 16(1971),
pp. 10.
48. Sato, M., Goto, M., Kondo, M., Kodama, A., Hirose, T. Countercurrent
extraction by supercritical carbon dioxide for the fractionation of citrus
oil. In: Proceedings of the 4th International Symposium on Supercritical
Fluids, 1114 May 1997, Sendai, Japan. Pp. 605608.
55
49. Schultz, W. G., Schultz, T. H., Carlson, R. A., Hudson, J. S. Pilot-plant
extraction with liquid CO2. Food Tech. 6(1974), pp. 32.
50. Hirohama, S., Takatsuka, T., Miyamoto, S., Muto, T. Phase equilibria for
the carbon dioxide-ethanol-water system with trace amounts of organic
components. J. Chem. Eng. Japan, 26(1993), pp. 247.
52. Simes, P. C., Matos, H. A., Carmelo, P. J., de Azevedo, E. G., Nunes da
Ponte, M. Mass transfer in countercurrent packed column : Application to
supercritical CO2 extraction of terpenes. Ind. Eng. Chem. Res., 34(1995),
pp. 613.
53. Carmelo, P. J., Pereira, P. J., Simoes, P. C., Nunes da Ponte, M. Scale-up
of a supercritical extraction unit for the deacidification of olive oil. In:
High Pressure Chemical Engineering, in Proceedings of the 3rd
International Symposium on High-Pressure Chemical Engineering, 79
Oct. 1996, Zrich, Switzerland. Pp. 487492.
54. Bernad, L., Keller, A., Barth, D., Perrut, M. Separation of ethanol from
aqueous solutions by supercritical carbon dioxide - comparison between
simulations and experiments. J. Supercritical Fluids, 6(1993), pp. 9.
56
57. Lim, J. S., Lee, Y.W., Kim, J.D., Lee, Y. Y. Mass-transfer and
hydraulic characteristics in spray and packed extraction columns for
supercritical carbon dioxide-ethanol-water system. J. Supercritical Fluids,
8(1995), pp. 127.
58. Nagase, Y., Tada, T., Ikawa, N., Fukuzato, R. Development of new
process of purification and concentration of ethanol solution using
supercritical carbon dioxide. In: Proceedings of the 4th International
Symposium on Supercritical Fluids, 1114 May 1997, Sendai, Japan. Pp.
617619.
57
64. Andrews, A.T. Supercritical Carbon Dioxide Extraction of Polycyclic
Aromatic Hydrocarbons from Contaminated Soil, PhD Dissertation,
Rutgers The State University of New Jersey - New Brunswick, 1990.
65. Treybal, R. E. Liquid Extraction. 2nd ed. McGraw-Hill, New York, 1963.
58
Published by Series title, number and
report code of publication
Vuorimiehentie 5, P.O.Box 2000, FIN-02044 VTT, Finland
Phone internat. +358 9 4561 VTT Publications 403
Fax +358 9 456 4374 VTTPUBS403
Author(s)
Laitinen, Antero
Title
Supercritical fluid extraction of organic compounds
from solids and aqueous solutions
Abstract
In this thesis, the extraction of organic compounds from solid and liquid matrices using supercritical carbon
dioxide as a solvent was studied.
Dense carbon dioxide is a very good solvent for the cyclic amide, 6-caprolactam. 6-caprolactam exhibits
relatively high solubilities, i.e. up to around 17 wt% at ambient temperatures, and at pressures below 220
bar. Carbon dioxide can be used to dissolve and extract 6-caprolactam from solid and liquid matrices.
The remediation of contaminated soil by carbon dioxide extraction was studied. The effect of extraction
pressure, temperature, soil moisture content, pH, carbon dioxide mass flow rate, and extraction time on the
residual pollutant concentration in the soil after extraction was systematically studied. The selected
pollutants were phenanthrene, 2,3,4,6-tetrachlorophenol and pentachlorophenol. The results indicate that at
least 8090 % of the initial amount of pollutant can be removed at moderate temperatures even in the most
difficult cases, and in many cases over 99 % of the initial amount of pollutant can be removed from the
soil. Supercritical extraction is best suited to sand or silt type soils, which have a low adsorption capacity.
The extraction results are found to be case-dependent, and soil remediation at moderate temperatures is
observed to be difficult if the pollutant binds strongly to the soil or if the acceptable level of residual
concentration is very low.
A novel mechanically agitated countercurrent continuous extraction column was introduced by extracting
ethanol and 1-butanol from aqueous solutions in the rotating disk column and in the Oldshue-Rushton
column. Column efficiencies were measured as a function of solvent-to-feed ratio and rotor speed.
Generally, the values of the height equivalent to a theoretical stage HETS were in the range of 0.4-0.8 m.
Mechanical agitation enhanced the extraction only at relatively high solvent-to-feed ratios. Column
capacities for both types of column were measured, and the total throughput calculated as the feed plus
solvent flow rate divided by the smallest cross-sectional area was 50-80 m3h-1m-2. A mechanically
agitated column successfully combines the high capacity of supercritical spray columns and the high
efficiency of supercritical packed columns.
Keywords
organic compounds, extraction, solids, aqueous solutions, supercritical gases, supercritical fluids, carbon
dioxide, amides, lactams, solubility
Activity unit
VTT Chemical Technology, Process Technology, Biologinkuja 7, P.O.Box 1401, FIN02044 VTT, Finland
ISBN Project number
9513855511 (soft back ed.) K9SU00433
951385552-X (URL: http://www.inf.vtt.fi/pdf/
Date Language Pages Price
December 1999 English 58 p. + app. 84 p. C
Name of project Commissioned by
National Technology Agency (Tekes)
Series title and ISSN Sold by
VTT Publications VTT Information Service
12350621 (soft back ed.) P.O.Box 2000, FIN02044 VTT, Finland
14550849 (URL: http://www.inf.vtt.fi/pdf/) Phone internat. +358 9 456 4404
Fax +358 9 456 4374